The Evidence for Acute Internal Medicine and Acute Medical Units

Future Hospital Journal 2016 Vol 3, No 1: 45–8 GENERALISM SPECIAL

T h e e v i d e n c e f o r a c u t e i n t e r n a l m e d i c i n e a n d a c u t e medical units

Author: M i k e J o n e sA

Acute medicine and acute medical units are relatively new recognised as a separate specialty with defined training 10 innovations. The evolving evidence base is demonstrating programmes. This organisation and specialisation means the effectiveness of these in improving care given to patients that the large majority of physician trainees now receive much with acute medical illness. This article reviews the available of their training in the care of acutely unwell medical patients evidence. while working in the AMU environment. ABSTRACT The key features and objectives of AIM are: KEYWORDS : Acute medicine , acute medical units > clinical service delivered by consultant physicians within the AMU environs for at least 12 hours per day and seven days per week Introduction > medical leadership within an AMU in a service designed and Acute medical units (AMUs) have been defined by the Royal managed by consultant physicians who specialise in AIM College of Physicians (RCP) as ‘a dedicated facility within a > service delivered by an AIM specialist and dedicated hospital that acts as the focus for acute medical care for patients multidisciplinary team with timely and appropriate who have presented as medical emergencies to hospital’.1 I n t h e interaction with other specialties past decade AMUs, as a base for the practise of acute internal > service design informed by quality standards medicine (AIM), have become integral to the care pathway of > performance benchmarked against key clinical quality most patients who require hospital-based acute medical care in indicators for AMU. 2,3 the UK. Other countries are increasingly adopting this model While surveys of care delivered within AMUs in the UK have 4 5,6 of care, including Ireland, Australasia and other parts of consistently reported heterogeneity with regard to organisation, 7–9 Europe. services and staffing,11–15 specific recommendations suggest AMUs first emerged in the 1990s often as a result of the local that medical care provided for patients in the AMU should service recognising the need to improve acute medical care. include the following. 1 However, there was limited evidence of the effectiveness of this model of care at that time. This review will discuss some of the 1. Patient assessment both in an ambulatory emergency care evidence which supports the acute medical model of care for setting and as part of a possible admission process patients admitted on the medical take. 2. Treatment – first-line treatments are often commenced within the short stay area (48–72 hours) of the AMU. 3. Further care planning – decisions about further Acute medical unit standards of care investigations and specialty referrals are made. Discharge In the early 1990s it was recognised that the established model planning should commence at the time of admission or of medical patients being admitted from the emergency a specialty bed requested. Effective multiprofessional department (ED) directly to a ward bed was unsustainable teamworking is integral to optimal AMU functioning. in the face of reduced bed capacity and increasing patient Where required, complex ethical decisions are usually made attendances/admissions. Alternative models of care were in the AMU, such as cardiopulmonary resuscitation status required. AIM evolved to provide patients with acute medical or deprivation of liberty orders. illness with the best quality care, in the right environment, 4. Practical procedures – a number of investigative and with assessment, diagnosis and treatment as actively managed therapeutic procedures are provided in an AMU. components of that care. Investigations such as lumbar puncture should be available The specialty of AIM was first introduced as a branch of in every AMU. The provision for other procedures is general internal medicine (GIM) but by 2009 was legally variable, such as echocardiography and ultrasound. 5. Delivery of high profile quality initiatives – there are a number of conditions where poor care is recognised in contributing to adverse patient outcomes. The highest Author: A consultant acute physician, Acute Medicine, University profile examples are sepsis and acute kidney injury. Hospital of North Durham, Durham, UK Both are seen frequently in AMUs and initiatives have

FHJv3n1-Jones.indd 45 20/01/16 12:05 PM Mike Jones

been undertaken to improve the care delivered for these transfers from an AMU following admission may introduce conditions. further delay. 6. Higher level care – an AMU does not provide high- dependency care but in some AMUs enhanced levels of care, M o r t a l i t y such as non-invasive ventilation or increased monitoring, may be provided. Mortality data presented in several studies were assessed at 7. Generic care – AMU is frequently the starting point for a variety of time points: in-hospital, 30-day post admission, 7,16–19,23,24,26,27 the care described in local commissioning for quality and 30-day post discharge and annual. I n m a n y o f innovation payments, eg assessment for and provision of the studies, patient admission to the AMU was associated VTE prophylaxis. with reduced mortality by comparison with patients 8. Access to diagnostic services – every AMU must be admitted under non-AMU models of care. By contrast, one supported by a full range of diagnostic investigations. study has reported non-significant increases in 30-day and 9. Access to higher level care – every AMU must be supported in-hospital mortality in those receiving AMU care. It is by a critical care unit and coronary care unit. notable however that in this study baseline mortality rates 7 10. Specialty in-reach – given the variety of patient were higher than those in comparable studies. In these nine presentations medical specialty in-reach or co-location studies the absolute change in mortality between the AMU from cardiology, medicine for care of the elderly and and non-AMU groups ranged from +1.3% to –8.8%. Four respiratory medicine is an absolute requirement. Ready studies attempted to adjust for confounding factors. The first availability of advice and specialty management pathways demonstrated that while not all patients could be allocated from the other medical specialties is critical. As well as to the AMU due to capacity constraints, the univariate medical specialities, AIM needs to work closely with other odds ratio of an in-hospital death by day 30 for a patient disciplines, for example surgical specialities, obstetrics and initially allocated to the AMU, compared with an initial gynaecology, intensive care medicine and psychiatry. ward allocation was 0.76 (95% CI 0.71, 0.82; p<0.001). The fully adjusted risk for patients was 0.67 (95% CI 0.62, 0.73; p<0.001). 16 A second study using propensity score matching The evidence for AMU care demonstrated a non-significant reduction in in-hospital Over the past decade, evidence has increasingly demonstrated mortality in the AMU cohort compared with the non-AMU the benefit of AMU models of care. Initial research has focused cohort (unmatched analysis 3.7 vs 4.6%; matched analysis 4.2 on the following outcomes as important quality indicators: vs 4.6%).17 Investigators undertaking a third study adjusted mortality, hospital length of stay (LOS), hospital readmission for time based changes in mortality in the population. In and patient/staff satisfaction. These will be discussed in more addition to an overall trend towards reduced mortality detail here. implementation of specialty triage was associated with a significant reduction in the subsequent mortality of the under Hospital length of stay 65 age group by a further 0.64% (95% CI 0.11, 1.17%; p = 0.021), equivalent to approximately 51 fewer deaths per year. Hospital LOS represents a composite measure which not There was, however, no significant effect of specialty triage for only indicates effectiveness of clinical treatment but also those aged over 65 or all age groups combined.18 A final study the efficiency of processes of care. Hospital LOS has been used logistic regression to adjust for confounding factors 7,8,16–26 assessed in many recent studies. The majority of including comorbidities, illness severity score and disease studies demonstrated a statistically significant difference in category. A significant reduction in in-hospital mortality in favour of AMU care. In the studies that reported the mean the AMU group by comparison with the non-AMU group was LOS for the two groups, the reduction ranged from 0.3 reported (adjusted odds ratio 0.28; 95% CI 0.23, 0.35).27 I n to 2.62 days. However, only three of these attempted to adjust general, these studies suggest that the admission of patients to for confounding factors. AMU in acute medical illness decreases overall mortality. In the first, which used propensity score matching, the mean reduction in LOS was 0.8 days in the matched analysis versus Hospital readmission 0.11 days in the unmatched. Both studies showed a statistically significant reduction in length of stay.17 A second study in Hospital readmissions following discharge are rightly which adjustment was made for secular trends, demonstrated considered an adverse outcome. While a failure of social and that the LOS was 0.73 days less in the AMU group when other community care may contribute to readmissions rates, compared to the non-AMU group (95% confidence interval such events are nevertheless regarded as markers of suboptimal (CI) –1.5, 0.04; p = 0.067). 18 A final study used multiple linear healthcare delivery. The studies quoted above have evaluated regression analysis to demonstrate that patients being cared the change in the proportion of patients readmitted to hospital. for entirely in the AMU were found to have a mean LOS 5.7 These were again measured at varying time points: 7, 28 or 30 days less than patients being cared for entirely on the ward days.7,17–19,25,26 Review of seven-day readmission demonstrated (p<0.001). 19 This, however, was not the case when the ward a non-significant reduction in the proportion of patients group was compared with those patients first treated in the readmitted in the AMU group compared to the non-AMU AMU and then transferred to the ward, with the latter group group (range 0.1– 0.7%).17,18,25 For patients readmitted within having a mean LOS just under a day longer than the ward group 28 or 30 days, three studies found a decrease in the AMU group (p = 0.04). In general, these studies suggest that hospital LOS compared to the non-AMU group 7,17,18 (range 0.7–5.2%) with benefits from the presence of an AMU, but by contrast, that only one reaching statistical significance. Only one study took

FHJv3n1-Jones.indd 46 20/01/16 12:05 PM Acute internal medicine and acute medical units

measures to control for confounding factors. Using a time series The future for the AMU will almost certainly see it evolve not analysis to compare summary changes between non-AMU only as the principal site for medical admissions but also the and AMU groups, investigators demonstrated no significant main community-facing service providing integration of care difference in either 7-day readmission (summary change with primary and social services. First, to manage and perhaps –0.02; 95% CI –0.07, 0.03; p = 0.365) or 28-day readmission prevent admissions through the development of adequate rates (–0.04, 95% CI –0.15, 0.07; p = 0.49).18 T h e r e f o r e , t o patient assessment prior to admission, ambulatory care clinics date there are no significant data demonstrating that hospital and the coordination of specialist hot clinics but further, to readmissions are decreased when an AMU is involved in acute facilitate discharge and follow-up care planning, and prevent medical patient care. readmission. By comparison with most other UK specialities, AIM is yet Patient/staff satisfaction in its infancy. Data, however, from investigations in the UK and elsewhere collected over the past five to ten years and Patient experience and satisfaction are increasingly seen as presented here suggest that there are significant benefits for important indices of excellent care but also correlate with better patients admitted under this model of care. Patient mortality clinical outcomes. Four recent studies have reviewed patient and LOS, but also patient and staff experience outcomes, 20,28–30 satisfaction in the context of the AMU model. all show improvements. The evidence base requires further Hanlon and colleagues reported a statistically significant expansion and, indeed, the precise contribution of the increase in the percentage of patients feeling ready for various elements of the AMU to these outcomes is unclear discharge, and reported that staff had time to explain their and remains a topic for further research. In the meantime 28 treatment in the AMU group. it is clear that the AIM with increasing numbers of acute A second study revealed contrasting results in a retrospective physicians will be the most common site responsible for care analysis of the Adult Inpatient Survey performed in the NHS of patients with unscheduled acute medical illness. Striving 30 in England on a total of 17,182 patients. This study, however, for excellence in the care delivered to patients in these units compared the AMU group with short-stay elective patients or will be fundamental to the functioning of the hospital of the unscheduled non-medical admissions. The numbers reviewed future. ■ in the other two studies did not allow for statistical analysis. Three of these studies have also reported results of staff References satisfaction surveys. 20,28,29 Hanlon and colleagues again reported statistically significant positive and negative effects of 1 Acute Medicine Task Force . Acute medical care. The right person in the AMU model from surveys of a total of 50 consultants and the right setting – first time. London : RCP , 2007 . 190 nurses. 27 Of note, non-consultant staff were less concerned 2 Beckett DJ , Raby E , Pal S , Jamdar R , Selby C . Improvement in about losing track of patients but were more concerned about time to treatment following establishment of a dedicated medical admissions unit. Emerg Med J 2009 ; 26 : 878 – 80 . ‘blocked’ beds, while nursing staff felt that there was more time 3 McNeill GBS , Brahmbhatt DH , Prevost AT , Trepte NJB . What is the for health promotion. In general, there was a perception of effect of a consultant presence in an acute medical unit? Clin Med increased stress but better job satisfaction in the AMU model. 2009 ; 9 : 214 – 8 . In the other studies the majority of medical and nursing staff 4 Watts M , Powys L , Hora CO et al. Acute medical assessment units: 20 felt the AMU model to be better and 75% of GPs preferred the an efficient alternative to in-hospital acute medical care. Ir Med J AMU route for patients compared to the ED.28 2011 ; 104 : 47 – 9 . 5 Providence C , Gommans J , Burns A . Managing acute medical D i s c u s s i o n admissions: a survey of acute medical services and medical assessment and planning units in New Zealand. Intern Med J Prior to the adoption of AIM and the AMU model of care 2012 ; 42 : 51 – 6 . there were a large variety of systems used within the NHS to 6 McNeill GBS , Brand C , Clark K et al. Optimizing care for acute configure the pathway undertaken by patients admitted to medical patients: the Australasian Medical Assessment Unit Survey . Int Med J 2011 ; 41 : 19 – 26 . hospital with acute medical illness. Acute admissions were 7 Vork JC , Brabrand M , Folkestad L et al . A medical admission unit usually cared for by physicians who had trained in general reduces duration of hospital stay and number of readmissions . medicine as well as another specialty. However, the competing Dan Med Bull 2011 ; 58 : A4298 . demands of speciality medicine meant that the care of patients 8 D i e p e v e e n B A . Performance analysis and improvement at the acute with acute medical illness was often not the principle focus admissions unit of Máxima Medical Centre . Master of Science in of consultant physician workload. The development of AIM Operations Management and Logistics [dissertation]. Eindhoven : has occurred in tandem with the strengthening of governance Technische Universiteit Eindhoven, 2009 . systems and quality improvement processes within the NHS. 9 Realdi G , Giannini S , Fioretto P et al. Diagnostic pathways of the Reflecting these changes, the AMU has become an important complex patients: rapid intensive observation in an Acute Medical site for improvement in the care of patients with acute illness. Unit. Int Emerg Med 2011 ; 6 Suppl 1: 85 – 92 . 10 JRCPTB . Acute internal medicine curriculum . London : JRCPTB For the first time, specialists from AIM, working in concert website , 2009 (modified 2012) . with specialty colleagues, focus on the care of patients 11 Ward D. Acute medical care. The right person, in the right setting admitted through this pathway to ensure the delivery of high- first time: how does practice match the report recommendations? quality acute care. The majority of physician trainees gain Clin Med 2009 ; 9 : 553 – 6 . experience admitting unselected medical patients in the AMU 12 Jayawarna C , Atkinson D , Ahmed SV , Leong K . Acute medicine units: environment, thus there is significant overlap between AIM and the current state of affairs in the North-West of England . J R Coll GIM training. Physicians Edinb 2010 ; 40 : 201 – 4 .

FHJv3n1-Jones.indd 47 20/01/16 12:05 PM Mike Jones

13 Ilyas M , Zahid M , Roseveare C . Acute medical care. Clin Med 23 Moloney ED , Bennett K , Silke B . Effect of an acute medical admission 2010 ; 10 : 304 . unit on key quality indicators assessed by funnel plots . Postgrad Med J 14 Le Jeune I , Masterton-Smith C , Subbe CP , Ward D . “State of the 2007 ; 83 : 659 – 63 . Nation”–the Society for Acute Medicine’s Benchmarking Audit 24 Conway R , O’Riordan D , Silke B . Long-term outcome of an 2013 (SAMBA ’13) . Acute Med 2013 ; 12 : 214 – 9 . AMAU-a decade's experience. QJM 2014 ; 107 : 43 – 9 . 15 Wood I. Medical assessment units in the West Midlands region: a 25 St Noble VJ , Davies G , Bell D . Improving continuity of care in an nursing perspective. Accid Emerg Nurs 2000 ; 8 : 196 – 200 . acute medical unit: initial outcomes. QJM 2008 ; 101 : 529 – 33 . 16 Coary R , Byrne D , O’Riordan D et al . Does admission via an acute 26 Brand CA , Kennedy MP , King-Kallimanis BL et al. Evaluation of the medical unit influence hospital mortality? 12 years’ experience in a impact of implementation of a medical assessment and planning large Dublin hospital . Acute Med 2014 ; 13 : 152 – 8 . unit on length of stay . Aust Health Rev 2010 ; 34 : 334 – 9 . 17 Li JYZ , Yong TY , Bennett DM et al . Outcomes of establishing an 27 Rooney T , Moloney ED , Bennett K , O’Riordan D , Silke B . Impact acute assessment unit in the general medical service of a tertiary of an acute medical admission unit on hospital mortality: a 5-year teaching hospital . Med J Aust 2010 ; 192 : 384 – 7 . prospective study. QJM 2008 ; 101 : 457 – 65 . 1 8 M o o r e S , G e m m e l l I , A l m o n d S et al . Impact of specialist care on 28 Watts M , Powys L , Hora CO et al . Acute medical assessment units: clinical outcomes for medical emergencies . Clin Med 2006 ; 6 : 286 – an efficient alternative to in-hospital acute medical care. Ir Med J 93 . 2011 ; 104 : 47 – 9 . 19 Suthers B , Pickles R , Boyle M et al. The effect of context on per- 29 Hanlon P. Coping with the inexorable rise in medical admissions: formance of an acute medical unit: experience from an Australian evaluating a radicalreorganisation of acute medical care in a Scottish tertiary hospital . Aust Health Rev 2012 ; 36 : 320 – 4 . district general hospital . Health Bull (Edinb) 1997 ; 55 : 176 – 84 . 20 McLaren EH. Re-organising emergency medical admitting : the 30 Sullivan P , Harris ML , Bell D . The quality of patient experience Stobhill experience, 1992–1997 . Health Bull (Edinb) 1999 ; 57 : 108 – of short-stay acute medical admissions: findings of the Adult 17 . Inpatient Survey in England . Clin Med 2013 ; 13 : 553 – 6 . 21 Moloney ED , Smith D , Bennett K , O’Riordan D , Silke B . Impact of an acute medical admission unit on length of hospital stay, and emergency department ‘wait times’. QJM 2005 ; 98 : 283 – 9 . Address for correspondence: Dr M Jones, Acute Medicine, 22 Moloney ED , Bennett K , O’Riordan D , Silke B . Emergency depart- University Hospital of North Durham, North Road, Durham ment census of patients awaiting admission following reorganisa- DH1 5TW, UK. tion of an admissions process. Emerg Med J 2006 ; 23 : 363 – 7 . Email: [email protected]

FHJv3n1-Jones.indd 48 20/01/16 12:05 PM